AI Agent
An AI system that takes a goal and acts on it across multiple steps, calling tools, APIs, or other models to get there, rather than just answering one prompt. In crypto, agents often hold a wallet and transact on-chain.
Also known as: agent, AI agents, autonomous agent
An AI agent is a model wired up to do things, not just say things. A plain language model answers a prompt and stops. An agent takes a goal (“research this token and post a summary”), breaks it into steps, and works through them: calling a search API, reading the results, calling an inference endpoint to reason over them, then calling a posting API to publish. The model is the brain; the agent is the brain plus hands, plus a loop that keeps going until the goal is met or it gives up.
In decentralised AI, agents matter because they’re where the money moves. An agent that holds a wallet can pay for its own inference, buy data, settle with other agents, and earn fees, all on-chain. That turns a chatbot into an economic actor. Most of the agent-platform projects OYM covers (ElizaOS, Virtuals, Autonolas, OpenServ) are competing to be the framework those agents are built and launched on.
The honest question to ask of any agent project is whether the agents do something verifiable, or whether “agent” is a label on a thin wrapper around someone else’s model. Rob’s recurring test applies: how do you know, with provable certainty, that the agent does what it’s claimed to do? Real autonomy, on-chain settlement, and open tooling are the signals; a Telegram bot calling GPT behind a token is the anti-signal.